1. Field of the Invention
The present invention relates generally to imaging of three-dimensional (“3D”) volume data sets. More particularly, the present invention relates to fast visualization and analysis of very large 3D volume data sets using 3D computer graphics.
2. Related Art
Many fields of endeavor require the analysis and imaging of three-dimensional (“3D”) volume data sets. For example, in the medical field, a CAT (computerized axial tomography) scanner or a magnetic resonance imaging (MRI) device is used to produce a “picture” or diagnostic image of some part of a patient's body. The scanner or MRI device generates a 3D volume data set that needs to be imaged or displayed so that medical personnel can analyze the image and form a diagnosis.
Three-dimensional volume data sets are also used in various fields of endeavor relating to the earth sciences. Seismic sounding is one method for exploring the subsurface geology of the earth. An underground explosion or earthquake excites seismic waves, similar to low frequency sound waves that travel below the surface of earth and are detected by seismographs. The seismographs record the time of arrival of the seismic waves, both direct and reflected waves. Knowing the time and place of the explosion or earthquake, the time of travel of the waves through the interior can be calculated and used to measure the velocity of the waves in the interior. A similar technique can be used for offshore oil and gas exploration. In offshore exploration, a ship tows a sound source and underwater hydrophones. Low frequency (e.g., 50 Hz) sound waves are generated by, for example, a pneumatic device that works like a balloon burst. The sounds bounce off rock layers below the sea floor and are picked up by the hydrophones. In this manner, subsurface sedimentary structures that trap oil, such as faults, folds, and domes, are “mapped” by the reflected waves. The data is processed to produce volume data sets that include a reflection or seismic amplitude data value at specified (x, y, z) locations within a geographic space.
A 3D volume data set is made up of “voxels” or volume elements. Each voxel has a numeric value for some measured or calculated property, e.g., seismic amplitude, of the volume at that location. One conventional approach to generating an image of a 3D volume data set is to cross-section the 3D volume data set into a plurality of two-dimensional (“2D”) cross-sections or slices. The image of the 3D volume data set is then built as a composite of the 2D slices. For example, the image of the 3D volume data set is generated by stacking the 2D slices in order, back-to-front, and then composited into a complete image. The user sees the image being built layer by layer as the composite grows. Although the user can see the internal organization or structure of the volume as the composite image grows, the traditional slice and composite technique is typically slow, particularly when very large 3D volume data sets are being used. Additionally, the slice and composite technique clutters the user's field of view with extraneous information, and interferes with the user's ability to accurately visualize and interpret features inherent in the 3D volume data set.
Computer software has been developed specifically for imaging 3D seismic data sets for the oil and gas industry. Examples of such conventional computer programs include VoxelGeo, available from Paradigm Geophysical, Houston, Tex., SeisWorks and EarthCube, available from Landmark Graphics Corporation, and IESX, available from GeoQuest. Such conventional computer programs have numerous deficiencies that preclude a user from quickly and accurately visualizing and interpreting features inherent in a 3D seismic data set. Conventional computer programs for visualizing and interpreting 3D seismic data operate on the full 3D volume of seismic data. Consequently, every time a change is made, such as a change to the transparency or opacity settings, the full 3D volume of seismic data must be processed, and the image redrawn. Even when such programs are run on highly efficient graphics supercomputers, the delay or lag in re-drawing the image is perceptible to the user. For a 3D volume containing 500 megabytes of seismic data, it can take on the order of 30-45 seconds for conventional programs to re-draw the complete image (frame rate of 0.03 to 0.02 frames per second, respectively). During the 30-45 second delay time, the mind of the user loses focus on the feature of interest, making it difficult to completely and properly analyze the seismic data.
Some conventional 3D seismic interpretation programs provide the capability to visualize and interpret a piece of the full 3D volume of seismic data. The user identifies the coordinates of the selected piece via a menu command. An image of the selected piece is drawn. The selected piece can then be rotated, if desired, at that location. However, to look at a different piece of the full 3D volume of seismic data, such as to follow a geologic feature that has been tentatively identified, the image must be interrupted, a new location or coordinates for the different piece is entered, and a new image is drawn containing the different piece. The interruption in the displayed image makes it difficult for the user to visualize any continuity between the two pieces of the full 3D volume of seismic data that have been imaged. This impedes the user's ability to interpret and identify the geologic features that are inherent in the full 3D volume of seismic data. Additionally, even though only a piece of the full 3D volume of seismic data is being visibly displayed, conventional 3D seismic interpretation programs continue processing the full 3D volume of seismic data to draw the image, thereby slowing the display of the image to the user.
Conventional 3D seismic interpretation programs provide the capability to “auto pick” and identify points that satisfy a voxel selection algorithm. However, these programs typically iterate through the full 3D volume of seismic data to identify the points that satisfy the voxel selection algorithm. This is time consuming even on a high speed graphics supercomputer. Additionally, conventional 3D seismic interpretation programs do not provide the capability to directly delete from the collection of picked voxels. The only way to “eliminate” points from the collection of picked voxels using conventional 3D seismic interpretation programs is to repeatedly adjust the selection criteria for the voxel selection algorithm until the points to be eliminated fall outside of the selection criteria for the displayed points that satisfy the voxel selection algorithm. Each time the selection criteria is adjusted, the image must be interrupted. This iterative process is time consuming, and interferes with the visualization process of the user.
Thus, there is a need in the art for a system and method for imaging 3D volume data sets that overcomes the deficiencies detailed above. Particularly, there is a need for a system and method that re-draws images of large 3D volume data sets in response to user input at a rate sufficiently fast that the user perceives an instantaneous or real-time change in the image, without perceptible delay or lag. There is a need for a system and method that allows a user to interactively change the displayed image in a continuous manner, without interruption or perceptible delay or lag. Such a system and method would allow a user to more quickly and accurately interpret and identify features inherent in 3D volume data sets.